In vivo, dendritic cells can cross-present virus-like particles using an endosome-to-cytosol pathway.

Abstract

Recombinant parvovirus-like particles (PPV-VLPs) are particulate exogenous Ags that induce strong CTL response in the absence of adjuvant. In the present report to decipher the mechanisms responsible for CTL activation by such exogenous Ag, we analyzed ex vivo and in vitro the mechanisms of capture and processing of PPV-VLPs by dendritic cells (DCs). In vivo, PPV-VLPs are very efficiently captured by CD8alpha- and CD8alpha+ DCs and then localize in late endosomes of DCs. Macropinocytosis and lipid rafts participate in PPV-VLPs capture. Processing of PPV-VLPs does not depend upon recycling of MHC class I molecules, but requires vacuolar acidification as well as proteasome activity, TAP translocation, and neosynthesis of MHC class I molecules. This study therefore shows that in vivo DCs can cross-present PPV-VLPs using an endosome-to-cytosol processing pathway.

Full text

Page 1

In Vivo, Dendritic Cells Can Cross-Present Virus-Like
Particles Using an Endosome-to-Cytosol Pathway1
Vı ´ctor Gabriel Moro ´n,* Paloma Rueda,†Christine Sedlik,‡and Claude Leclerc2*
Recombinant parvovirus-like particles (PPV-VLPs) are particulate exogenous Ags that induce strong CTL response in the absence
of adjuvant. In the present report to decipher the mechanisms responsible for CTL activation by such exogenous Ag, we analyzed
ex vivo and in vitro the mechanisms of capture and processing of PPV-VLPs by dendritic cells (DCs). In vivo, PPV-VLPs are very
efficiently captured by CD8??and CD8??DCs and then localize in late endosomes of DCs. Macropinocytosis and lipid rafts
participate in PPV-VLPs capture. Processing of PPV-VLPs does not depend upon recycling of MHC class I molecules, but requires
vacuolar acidification as well as proteasome activity, TAP translocation, and neosynthesis of MHC class I molecules. This study
therefore shows that in vivo DCs can cross-present PPV-VLPs using an endosome-to-cytosol processing pathway. The Journal
of Immunology, 2003, 171: 2242–2250.
T
to target only Ags synthesized within the APC, for instance when
they are infected by a pathogen. Thus, Ags that do not reach the
cytosol of APCs should not elicit a CTL response. However, not
all pathogens are capable of infecting APCs, and moreover, this
mechanism cannot explain how CTL responses can be induced
against tumor cells. It is now well established that at least some
exogenous, noncytosolic Ags can gain access to the MHC class I
pathway of APCs by cross-presentation (1) through several alter-
native pathways of processing (2). Both macrophages (M?)3(3)
and dendritic cells (DCs) (4) cross-present Ags, but only DCs are
capable of stimulating naive CD8?T cells (5). Two main routes of
cross-presentation have been proposed. One route involves the
passage of Ags from endosomes to cytosol (also named cytosolic
diversion) (6), while in the other route the Ags do not escape from
endosomes, but are processed inside these vesicles (7, 8). The first
route seems to be used mostly by DCs, and the second one by M?
(9). A third route, which uses the properties of certain viruses and
he priming of T cell responses relies on the presentation
of Ag-derived peptides by MHC molecules. MHC class
I-restricted T cell responses were, until recently, thought
bacterial toxins to gain direct access to cytosol, is based on the
translocation of Ag from cell surface to cytosol (10, 11).
We have developed an Ag delivery system based on nonrepli-
cative, recombinant parvovirus-virus-like particles (PPV-VLPs)
formed by the self-assembly of the VP2 capsid protein of porcine
parvovirus (PPV) (12, 13). The VP2 protein (14), carrying foreign
CD8?T cell epitopes, self-assembles into 25-nm VLPs after ex-
pression in insect cells (13). Mice immunized with PPV-VLPs
carrying a CD8?T cell epitope, in the absence of an adjuvant,
develop a strong and specific MHC class I-restricted CTL response
(13, 15). This CTL response protects against a viral challenge and
is based on the induction of a high frequency of high avidity CTLs
(16). PPV-VLPs target DCs with very high efficiency, whereas M?
and B cells have a poor capacity to capture these particles. Fol-
lowing an in vivo injection of PPV-VLPs, both CD8??and
CD8??DCs capture and process these particles, although they
have different kinetics and T cell requirements (15). DC stimula-
tion by PPV-VLPs induces phenotypic changes on CD8??DCs,
leading to the acquisition of CD8? and CD205 and the loss of CD4
molecules as well as expression of costimulatory molecules on
both DC subsets (15).
Most studies that aimed at studying the different steps of pro-
cessing of exogenous Ags into the MHC class I pathway have been
conducted using unpurified peritoneal exudate cells (17) (which
contain various cell populations, including M? and DCs) or bone-
marrow derived DCs (18–20), which do not represent the whole
spectrum of peripheral DCs. In the present report to decipher the
mechanisms responsible for CTL activation by exogenous Ags, we
analyze ex vivo and in vitro the mechanisms of capture and process-
ing of PPV-VLPs by DCs, using DCs freshly purified from spleen.
This study shows that DCs can use the endosome-to-cytosol pathway
in vivo to cross-present exogenous Ags such as PPV-VLPs.
Materials and Methods
Mice
Six- to 8-wk-old female C57BL/6 (H-2b) mice were obtained from Janvier
(Le Genet St. Isle, France). Female TAP1 knockout mice, bred onto a
C57BL/6 background, were a gift from Dr. A. Bandeira (Institut Pasteur,
Paris, France). All animals were maintained under specific pathogen-free
conditions.
*Unite ´ de Biologie des Re ´gulations Immunitaires, Institut National de la Santé et de
la Recherche Médicale, E352, Institut Pasteur, Paris, France;†Inmunologı ´a y Ge-
ne ´tica Aplicada S.A., Madrid, Spain; and‡Institut National de la Sante ´ et de la Re-
cherche Me ´dicale Unite ´ 520, Institut Curie, Paris, France
Received for publication December 31, 2002. Accepted for publication June 24, 2003.
The costs of publication of this article were defrayed in part by the payment of page
charges. This article must therefore be hereby marked advertisement in accordance
with 18 U.S.C. Section 1734 solely to indicate this fact.
1This project is a collaborative work between Pasteur Institute and INGENASA,
supported by European Commission Grant QLK2-CT-1999-00318. G.M. was sup-
ported by a postdoctoral fellowship from the Consejo Nacional de Investigaciones
Cientı ´ficas y Tecnolo ´gicas (Argentine), European Economic Community Grant
QLK2-CT-1999-00429, and the Fondation de la Recherche Medicale (France). C.S.
was supported by grants from the European Union (Contract BMH4-CT 98-3703).
2Address correspondence and reprint requests to Dr. Claude Leclerc, Departement
d’Immunologie, Institut Pasteur, 25 rue du Docteur Roux 75724, Paris Cedex 15,
France. E-mail address: cleclerc@pasteur.fr
3Abbreviations used in this paper: M?, macrophages; BFA, brefeldin A; CCB, cy-
tochalasin B; CHX, cycloheximide; DC, dendritic cell; DMA, dimethylamiloride; ER,
endoplasmic reticulum; LLmL, N-acetyl-L-leucinal-L-methioninal; LLnL, N-acetyl-L-
leucinal-L-norleucinal; OVA257–264, a Kb-restricted epitope corresponding to aa 257–
264 of chicken egg albumin; PPV-VLP, porcine parvovirus-virus-like particle; PPV-
VLPs-OVA, porcine parvovirus-virus-like particles carrying the OVA257–264epitope;
WGA, wheat-germ agglutinin; VLP, virus-like particle; HBsAg, hepatitis B virus
small envelope protein.
The Journal of Immunology
Copyright © 2003 by The American Association of Immunologists, Inc.0022-1767/03/$02.00

Page 2

Peptides and cell lines
The peptide SIINFEKL, corresponding to aa 257–264 from chicken OVA
(OVA257–264peptide), an immunodominant H-2b-restricted CTL epitope,
was purchased from Neosystem (Strasbourg, France). B3Z, a CD8?T cell
hybridoma specific for the OVA257–264epitope in the context of Kb(21),
was a gift from Dr. N. Shastri (University of California, Berkeley, CA).
PPV particles
The construction, characterization, and purification of recombinant and
control PPV-VLPs were previously described in detail (13, 15). Briefly, the
VP2 gene was expressed with the OVA257–264peptide plus natural flanking
sequences (LEQLESIINFEKLTE) (the underlined sequence corresponds to
the CTL epitope, and the nonunderlined sequences correspond to flanking
sequences to the CTL epitope) from chicken egg OVA in its 5? end (PPV-
VLPs-OVA) or without this sequence (PPV-VLPs) using a baculovirus
vector system. After infection of Sf9 insect cells, the recombinant VLPs
were purified by salt precipitation with 20% ammonium sulfate, followed
by dialysis. Characterization of PPV-VLPs-OVA and PPV-VLPs by CsCl
sedimentation analysis and electron microscopy revealed properties iden-
tical to those of native PPV virions. In some experiments PPV-VLPs-OVA
were labeled with the fluorochrome Alexa488, using the Alexa Fluor 488
Protein Labeling Kit (Molecular Probes Europe, Leiden, The Netherlands)
following the manufacturer’s instructions.
The concentration of PPV-VLPs-OVA was determined by densitometry
and by double-Ab sandwich ELISA. The densitometric assay was con-
ducted with 1D Image Analysis software 2.0.1. (Eastman Kodak, Roches-
ter, NY) using BSA as reference. The double-Ab sandwich ELISA was
performed as previously described (22), using as capture Ab the anti-PPV
mAb 15C5 and as detection Ab the anti-PPV biotinylated mAb 13C6 (23).
Highly purified PPV-VLPs from size exclusion chromatography were used
as the standard reference.
Endotoxin values were determined in each sample of VLPs using the
Limulus amebocyte lysate test (BioWhittaker, Walkersville, MD). For
PPV-VLPs, endotoxin values were ?0.5 ng/mg of protein (5 endotoxin
units/mg), and for PPV-VLPs-OVA, they were ?10 ng/mg (100 endotoxin
units/mg). PPV-VLPs preparations contained minimal traces of DNA, as
determined after DNAzol (Invitrogen, Paisley, U.K.) or phenol extractions
and electrophoresis in agarose gel, although not enough to be quantified by
spectrophotometry at A260/280.
CyaA-E5-OVA
CyaA-E5-OVA is a genetically detoxified CyaA toxin carrying the
OVA257–264peptide plus natural flanking sequences (PASIINFEKLGT)
between Arg224and Ala225of the amino acid sequence (11).
Preparation of splenic DCs
Spleens were removed from mice and treated for 45 min at 37°C with 400
U/ml of collagenase type IV and 50 ?g/ml of DNase I (Roche, Mannheim,
Germany) in RPMI 1640. After inhibition of collagenase activity with 6
mM EDTA in PBS, spleens were dissociated in Ca2?- and Mg2?-free PBS
in the presence of 2.5 mM EDTA and 0.5% FCS (Life Technologies, Pais-
ley, Scotland). In all assays involving DCs, the same batch of endotoxin-
free FCS (as determined by Limulus amebocyte lysate test) was used. Fur-
thermore, all reagents were also tested for endotoxins. Single spleen cell
suspensions were prepared and incubated with anti-CD16/32 (2.4G2 clone;
BD PharMingen, San Diego, CA) and with colloidal superparamagnetic
microbeads conjugated to anti-CD11c mAb (MACS-anti-CD11c, N418
clone; Miltenyi Biotec, Bergisch-Gladbach, Germany) following the man-
ufacturer’s instructions. CD11c?cells were positively selected with high
speed magnetic cell sorting (program posseld, AutoMACS; Miltenyi Bio-
tec). The purified DC preparations contained 3–10% autofluorescent cells
(defined as double-positive cells in a FL2 vs FL3 dot plot, without Ab
labeling). The purity of DC preparations (excluding autofluorescent cells)
was always 95–99% (Fig. 1A). CD11c?cells were H-2 Kb?, I-Ab low,
CD40low, CD80low, and CD86?. Twenty-five to 30% were CD8??, and
60–70% were CD8??CD11b?.
Ag presentation assay
CD11c?spleen cells (105cells/well) were first pulsed with Ag (PPV-
VLPs-OVA, PPV-VLPs, or OVA257–264peptide) for 4 h in 96-well culture
microplates in a final volume of 0.2 ml of RPMI 1640 Glutamax-I plus 5 ?
10?5M 2-ME, 100 IU/ml penicillin, 100 ?g/ml streptomycin, and 10%
FCS (RPMI 10%; all from Life Technologies). The Ag concentration used
in each experiment is indicated in the figure legends. Then DCs were
washed three times with RPMI 10%, and 105cells/well of B3Z hybridoma
were added and incubated overnight at 37°C in 95% CO2. The stimulation
of B3Z cells was monitored by IL-2 release in the supernatants, which was
measured using the CTLL-2 bioassay. Cells (104/well) of the CTLL-2 cell
line were cultured with 100 ?l of supernatant in a final volume of 200 ?l.
Two days later, [3H]thymidine (NEN Life Science, Boston, MA) was
added, and the cells were harvested 18 h later with an automated cell
harvester (Skatron, Lier, Norway). Incorporated thymidine was detected by
cell scintillation counting. In all experiments each point was determined in
duplicate. Results are expressed as counts per minute. In some experiments
APCs were fixed before or after Ag pulse with 0.05% glutaraldehyde/PBS,
then treated with 0.2 M lysine/RPMI and washed three times with
RPMI 10%.
Inhibition studies
For most inhibition studies, APCs (105cells/well) were first incubated in
0.1 ml with each drug for 1 h. Then Ag diluted in 0.1 ml was added to the
wells (0.2 ml final volume) at the final concentration indicated in the con-
tinuous presence of inhibitors for 4 h. APC were then washed three times
and fixed with 0.05% glutaraldehyde, and the experiments were continued
as described above. The inhibitors used were brefeldin A (BFA), N-acetyl-
L-leucinal-L-norleucinal (LLnL), N-acetyl-L-leucinal-L-methioninal (LLmL),
dimethylamiloride (DMA), chloroquin, cytochalasin B (CCB), primaquin, cy-
cloheximide (CHX), chlorpromazine, filipin III (all from Sigma-Aldrich, St.
Louis, MO), lactacystin (BIOMOL, Plymouth Meeting, PA), leupeptin, and
pepstatin (Roche, Indianapolis, IN).
For inhibition of clathrin-mediated endocytosis by K?depletion follow-
ing hypotonic shock, DCs (105cells/well) were preincubated for 30 min in
the presence of serum-free synthetic OptiMEM medium (Life Technolo-
gies), supplemented with 5 ? 10?5M 2-ME, 100 IU/ml penicillin, and 100
?g/ml streptomycin. DCs were then incubated for 5 min in hypotonic me-
dium (OptiMEM mixed with Ultrapure H2O, 50/50) and finally for 30 min
in isotonic K?-free (140 mM NaCl, 20 mM HEPES-NaOH, 1 mM CaCl2,
1 mM MgCl2, 1 mg/ml glucose, and 0.5% BSA) or K?-containing (10 mM
KCl, 130 mM NaCl, 20 mM HEPES-NaOH, 1 mM CaCl2, 1 mM MgCl2,
and 0.5% BSA) buffer. Then 1.23 nM PPV-VLPs-OVA or 1.1 nM
FIGURE 1.
DCs, delivering the OVA257–264epitope into the MHC class I pathway. A,
One C57BL/6 mouse was i.v. injected with 50 ?g of Alexa488-PPV-VLPs.
One PBS-injected mouse was used as a control. Ninety minutes later,
CD11c?cells were purified from spleen by magnetic sorting, stained with
PE-anti-CD11c, and analyzed on a FACStar cytometer. A representative
dot-plot analysis of cells purified from PPV-VLP-injected or control mice
is shown. B, C57BL/6 mice were i.v. injected with 100 ?g of PPV-VLPs-
OVA (F) or PPV-VLPs (E), and 90 min later, various numbers of DCs
purified from their spleens were cocultured overnight with 1 ? 105B3Z
cells/well. The presentation of OVA257–264peptide to B3Z cells was mon-
itored by IL-2 production, measured by a CTLL-2 proliferation assay, and
expressed as the mean ? SEM counts per minute of duplicate wells. C,
Purified splenic DCs (105cells/well) were incubated in vitro with PPV-
VLPs-OVA (F), control PPV-VLPs (f), or OVA257–264peptide (E) for
4 h, washed, and cultured overnight with 1 ? 105cells/well of B3Z cells.
The presentation of OVA257–264peptide to B3Z cells was monitored by
IL-2 production, measured by a CTLL-2 proliferation assay, and expressed
as the mean ? SEM counts per minute of duplicate wells.
In vivo, PPV-VLPs-OVA are captured and presented by
2243The Journal of Immunology

Page 3

OVA257–264were added. DCs were incubated for 1 h in the continued
presence of the drugs and Ags. Then DCs were washed and cocultured
overnight with 1 ? 105cells/well of OVA-specific B3Z cells in RPMI 10%
. As control, cells were maintained in OptiMEM medium and incubated
with Ag before washing and culture with B3Z cells in RPMI 10% as in-
dicated above.
Confocal microscopy and in vitro internalization assay
CD11c?spleen cells from PPV-VLP-injected mice were allowed to adhere
to glass slides coated with poly-L-lysine (Sigma-Aldrich). Cells were then
fixed with 4% paraformaldehyde and in some cases labeled with wheat-
germ agglutinin (WGA; Molecular Probes Europe,) directly conjugated to
Alexa Fluor 488 for 30 min without permeabilization of the cells to detect
plasma membrane. Intracellular immunofluorescence was performed on
cells permeabilized in PBS containing 1% saponin and 5% BSA with Abs
against CD8? (53-6.7 clone; BD PharMingen), Lamp2 (BD PharMingen),
H2-M (DM5 clone) (24), Rab7 (supplied by P. Chavrier, Institut Curie,
Paris, France) (24, 25), Rab11 (26), or biotinylated anti-VP2 Ab (mixture
of 11D1, 13C4, 13C5, 13C6, and 15G4 mAbs) (23). After incubation with
the appropriate secondary Abs, cells were mounted and analyzed by con-
focal microscopy using a Leica TCS SP2 microscope (Leica, Deerfield, IL)
equipped with a ?100 1.4 NA HCX PL APO oil immersion objective.
For in vitro internalization assay, CD11c?spleen cells from noninjected
mice were incubated with Alexa488-PPV-VLPs for 1 h at 37°C, followed
by rhodamine-WGA (Molecular Probes Europe) staining as described
above.
Results
In vivo and in vitro, DCs capture and process PPV-VLPs-OVA
To characterize the in vivo and in vitro processing of PPV-VLPs,
we first studied the in vivo capture of PPV-VLPs by DCs. Naive
C57BL/6 mice were injected with PPV-VLPs-OVA coupled to
Alexa488, a strong green fluorescent dye, which fluorescence does
not extinguish at low pH. Alexa488-PPV-VLPs-OVA preserved
their biological activity after labeling, as demonstrated by their
capacity to be processed and presented by DCs to B3Z cells (data
not shown). Ninety minutes after injection, the CD11c?spleen
cells were sorted out and labeled with an anti-CD11c mAb. As
shown in a representative experiment depicted in Fig. 1A, 53% of
CD11c?cells were Alexa488?showing that in vivo, DCs cap-
tured PPV-VLPs-OVA very efficiently. Fifteen hours after injec-
tion, DCs remained strongly positive for Alexa488 (52% of Al-
exa488?DCs; data not shown).
We then examined whether DCs can process PPV-VLPs in vivo,
using an ex vivo Ag presentation assay. C57BL/6 mice were i.v.
injected with 50 ?g (13 pmol/mouse) of PPV-VLPs-OVA or PPV-
VLPs. Ninety minutes later, splenic CD11c?cells were purified
and cocultured with B3Z CD8?T cell hybridoma. DCs purified
from PPV-VLPs-OVA-injected mice were capable of presenting
the OVA257–264epitope to B3Z cells, whereas DCs purified from
control PPV-VLP-injected mice failed to stimulate this hybridoma
(Fig. 1B). Thus, in vivo, DCs efficiently capture and process
PPV-VLPs-OVA.
To evaluate the capacity of DCs in vitro to process PPV-VLPs-
OVA, splenic CD11c?spleen cells purified from naive mice were
incubated with PPV-VLPs-OVA or PPV-VLPs or the OVA257–264
peptide for 4 h and then cocultured overnight with the B3Z hy-
bridoma. DCs incubated with PPV-VLPs-OVA or the OVA257–264
peptide efficiently presented the OVA257–264epitope, whereas DCs
incubated with PPV-VLPs did not stimulate B3Z cells (Fig. 1C).
Moreover, presentation of PPV-VLPs-OVA requires cellular pro-
cessing, because DCs fixed with glutaraldehyde before incubation
with PPV-VLPs-OVA did not stimulate B3Z. In contrast, DCs
fixed after PPV-VLPs-OVA incubation were fully capable of stim-
ulating this hybridoma (data not shown).
After capture, PPV-VLPs are localized in late endosomes of DCs
To analyze in detail the capture of PPV-VLPs by DCs, we studied
by confocal microscopy the intracellular localization of PPV-VLPs
FIGURE 2.
somes/lysosomes in DCs. A, Intracellular location of PPV-VLPs. Mice were
injected with 50 ?g of Alexa488-PPV-VLPs, and 90 min later, splenic DCs
werepurified.Cellswerethenstainedinthepresenceortheabsenceofsaponin
to permeabilize the cells, with biotinylated anti-PPV-VLPs mAbs detected
with Cy3-streptavidin. Far left figures show Alexa488 fluorescence, left im-
ages show Cy3 fluorescence, while right figures show the merged images. Far
right figures show the transmission image. B, Capture of PPV-VLPs by
CD8??and CD8??DCs. Mice were injected with 50 ?g of Alexa488-PPV-
VLPs, and 90 min later, splenic DCs were purified. Cells were then perme-
abilizedandlabeledwithananti-CD8?mAb(inred),whichwasdetectedwith
a Cy3-anti-rat secondary Ab. The transmission light image shows the body of
the cell. C, Mice were injected with unlabeled PPV-VLPs, and 90 min later,
splenic DCs were fixed, and different intracellular markers were identified by
specific antibodies detected with Alexa488-conjugated secondary Abs. Intra-
cellular location of PPV-VLPs was detected on permeabilized cells with anti-
PPV mAb, followed by Cy3-streptavidin. D, Purified splenic DCs from naive
mice were incubated with 5 ?g/ml of Alexa488-PPV-VLPs for 1 h at 37°C,
stained with rhodamine-WGA, and analyzed by confocal microscopy to visu-
alize the plasma membrane.
In vivo, captured PPV-VLPs-OVA are found in late endo-
2244CROSS-PRESENTATION OF VLPs BY DCs

Page 4

in DCs. Mice were injected with Alexa488-PPV-VLPs-OVA, and
their splenic DCs were purified, permeabilized or not with saponin,
and stained with a mixture of anti-VP2 mAbs. This mix of mAbs
colocalized with Alexa488 labeling in saponin-permeabilized DCs
from mice injected with Alexa488-PPV-VLPs, whereas no colo-
calization was observed in nonpermeabilized DCs from the same
mice (Fig. 2A), clearly showing that PPV-VLPs were effectively
found inside DCs and demonstrating that in vivo, DCs endocytose
PPV-VLPs. Furthermore, PPV-VLPs were found inside both
CD8??and CD8??DCs (Fig. 2B). Inside DCs, PPV-VLPs were
observed in vesicle-shaped structures of very different sizes (Fig.
2, A–C). Therefore, to characterize these vesicles, DCs from mice
injected with PPV-VLPs were stained with anti-VP2 mAbs and
with mAbs against several molecules recognized as markers of
endosomal compartments. The confocal microscopy revealed that
PPV-VLPs colocalized with Lamp2?, H2-M?, and rab7?labeling
(Fig. 2C), which are associated with late endosomes (27–29). Fur-
thermore, no colocalization of PPV-VLPs was observed with
rab11, a GTPase of the Rab family that is a crucial element in the
control of traffic through the recycling endosome (30, 31) (Fig.
2C). Therefore, after endocytosis, PPV-VLPs reach late endo-
somes. To determine whether this capture can be also observed in
vitro, purified splenic DCs from naive mice were incubated with
Alexa488-PPV-VLPs. The confocal analysis of DCs stained with
WGA (which recognizes (GlcNAc)2residues in cell surface gly-
coproteins) clearly showed that PPV-VLPs were effectively found
inside DCs (Fig. 2D), demonstrating that in vitro, DCs can also
endocytose PPV-VLPs. Moreover, in vitro endocytosed PPV-
VLPs were also found inside Lamp2?, H2-M?, and rab7?vesi-
cles, confirming the in vivo results (data not shown).
Capture of PPV-VLPs requires macropinocytosis, lipids rafts,
and actin polymerization, but not clathrin-coated pits
Endocytic pathways may include clathrin-coated pits, macropino-
cytosis, phagocytosis, or lipid rafts containing or not containing
caveolin. To define the pathways involved in PPV-VLP capture,
we used several drugs that are known to inhibit with a relative
specificity a particular step of these various pathways. Clathrin-
mediated endocytosis was assessed by K?depletion following hy-
potonic shock and by inhibition with chlorpromazine. K?deple-
tion following hypotonic shock was performed by cell exposure to
hypotonic medium, followed by incubation in the absence of ex-
tracellular potassium (32–34). This treatment results in dissocia-
tion of clathrin coats from the plasma membrane and nonproduc-
tive assembly of clathrin cages in the cytoplasm adjacent to coated
pits. As consequence, internalization is impaired for all membrane
proteins carrying cytoplasmic amino acid sequences that interact
with the clathrin adapter complex AP2. Upon this treatment, DCs
incubated with PPV-VLPs-OVA or OVA257–264peptide did not
show any inhibition of OVA257–264presentation (Fig. 3A).
As control, we used a second vector developed in our labora-
tory, the adenylate cyclase (CyaA) toxoid from Bordetella pertus-
sis (for review, see El Azami El Idrissi et al. (35)). CyaA is a
1706-aa protein that targets DCs, macrophages, neutrophils, and
NK cells through interaction with CD11b (11, 36, 37). Recombi-
nant CyaA toxin, carrying foreign CD8?T cell epitopes, can in-
duce a strong CTL response after injection in mice (38). In general,
CD11b binding is followed by the formation of clathrin-coated pits
(39). Indeed, spleen DCs K?-depleted following hypotonic shock
and coincubated with a recombinant CyaA containing the
FIGURE 3.
splenic DCs (105cells/well) were preincubated for 30 min in the presence of serum-free synthetic OptiMEM, then for 5 min in hypotonic medium and
finally for 30 min in isotonic K?-free or K?-containing buffer. Then, 1.2 nM PPV-VLPs-OVA, 27 nM CyaA-E5-MalE-OVA, or 1.1 nM OVA257–264was
added. DCs were then incubated for 1 h in the continued presence of the drugs and Ags. Then DCs were washed and cocultured overnight with 1 ? 105
cells/well of OVA-specific B3Z cells in RPMI 10% . As control, cells were incubated with OptiMEM and the above-indicated Ags before washing and
culture of B3Z cells. B, Purified splenic DCs (105cells/well) were preincubated for 1 h in the presence of the indicated concentrations of chlorpromazine.
Then 1.2 nM PPV-VLPs-OVA, 27 nM CyaA-E5-MalE-OVA, or 1.1 nM OVA257–264was added. DCs were incubated for 1 h in the continued presence
of the drugs and Ags. C–E, Purified splenic DCs (105cells/well) were preincubated for 1 h in the presence of the indicated concentrations of DMA (C),
filipin III (D), or 10 ?g/ml cytochalasin B (E). Then 1.2 nM PPV-VLPs-OVA or 1.1 nM OVA257–264was added. DCs were incubated for 4 h in the
continued presence of the drugs and Ags. The DCs were cocultured overnight with 105cells/well of B3Z-specific hybridoma. The presentation of
OVA257–264epitope to B3Z cells was monitored by IL-2 production and expressed as the percentage of [3H]thymidine incorporation compared with DCs
incubated with PPV-VLPs-OVA or OVA257–264peptide in the absence of inhibitor.
PPV-VLP uptake involves macropinocytosis, lipid rafts, and actin cytoskeleton polymerization, but not clathrin-coated pits. A, Purified
2245 The Journal of Immunology

Page 5

OVA257–264epitope (CyaA-E5-OVA) were incapable of present-
ing the OVA epitope to the hybridoma cells (Fig. 3A).
To confirm these results, we used chlorpromazine as an inhibitor
of clathrin-coated pit formation. Chlorpromazine is a cationic am-
phiphilic drug that has been extensively used to analyze the role of
clathrin-mediated endocytosis in virus entry (40, 41). OVA257–264
epitope presentation by DCs pretreated with chlorpromazine and
then incubated with PPV-VLPs-OVA was not altered, whereas
CyaA-E5-OVA presentation was reduced by 50% compared with
untreated control DCs (Fig. 3B). This result confirmed that clath-
rin-coated pits do not play a significant role in the capture of
PPV-VLPs.
To study whether PPV-VLP uptake is mediated by macropino-
cytosis, DCs were preincubated with DMA, which inhibits
Na?/H?exchange (42) and, therefore, macropinocytosis (43). As
shown in Fig. 3C, DMA strongly inhibited PPV-VLPs-OVA pre-
sentation, whereas it slightly affected the OVA257–264peptide pre-
sentation at a very high concentration of DMA.
Lipid rafts also participate in PPV-VLPs uptake, as revealed by
the use of filipin III inhibitor. Filipin III is a sterol-binding agent
that binds to cholesterol, disrupting lipid raft formation, including
caveolin-containing rafts, without affecting clathrin-mediated en-
docytosis (44). OVA257–264epitope presentation by DCs pre-
treated with filipin III and then incubated with PPV-VLPs-OVA
was severely reduced, whereas this treatment did not affect the
presentation of the OVA257–264peptide (Fig. 3D).
Finally, we studied PPV-VLPs processing by DCs preincubated
with CCB, a known inhibitor of actin filament polymerization (45),
which, in turn, can alter macropinocytosis, phagocytosis, as well as
caveolae-mediated endocytosis. DCs incubated with PPV-VLPs-
OVAinthepresenceofCCBwerenotabletopresenttheOVA257–264
epitope (Fig. 3E), whereas DCs incubated with the OVA257–264pep-
tide in the presence of CCB were fully able to activate B3Z cells.
The OVA257–264epitope processed from PPV-VLPs-OVA does
not bind to recycling MHC class I molecules
Most endocytosed molecules, including recycling receptors, are
delivered to early endosomes, where efficient sorting occurs (31).
Receptors and some ligands segregate to recycling compartments,
whereas other molecules are delivered into late endosomes-lyso-
somes. Recycling of endocytosed MHC class I molecules back to
the cell surface has been observed (46). Some of the recycling
MHC class I molecules can be loaded into endosomes with peptide
derived from endocytosed molecules (20, 47). Therefore, to estab-
lish whether PPV-VLP processing involves MHC class I recycling,
we studied PPV-VLPs-OVA presentation by DCs incubated in the
presence of primaquin, which inhibits the recycling of MHC class
I and II molecules (46). DCs incubated in the presence of this drug
presented efficiently both PPV-VLPs-OVA and the OVA257–264
peptide (Fig. 4A). This result is in accord with the absence of
colocalization of PPV-VLPs with rab11 by confocal microscopy
(Fig. 2C). Therefore, PPV-VLPs would not enter into recycling
compartments. In contrast, OVA257–264presentation was fully in-
hibited when DCs were incubated with PPV-VLPs-OVA in the
presence of cycloheximide (Fig. 4B), a potent inhibitor of protein
synthesis, suggesting that newly synthesized proteins, particularly
MHC class I molecules, are required for efficient presentation of
PPV-VLPs-OVA to B3Z cells.
PPV-VLPs processing needs vacuolar acidification and protease
hydrolysis
Purified DCs were incubated with PPV-VLPs-OVA in the pres-
ence of chloroquin, a known inhibitor of acidification of late en-
dosomes and lysosomes. Chloroquin strongly inhibited PPV-
VLPs-OVA presentation, without altering OVA257–264peptide
presentation, showing that acidification of late endosomes is nec-
essary for PPV-VLPs processing (Fig. 5A).
To analyze the participation of proteases in the processing of
PPV-VLPs, two different proteases inhibitors, leupeptin (which is
an inhibitor of serine and cysteine proteases, including cathepsin
B) and pepstatin (a inhibitor of aspartate proteases, including
cathepsins D and E) (48, 49), were used. Incubation of DCs with
PPV-VLPs in the presence of leupeptin did not block PPV-VLP-
OVA presentation (Fig. 5B), whereas it inhibited the MHC class
II-restricted presentation of exogenous proteins (data not shown).
Incubation in the presence of pepstatin resulted in a partial inhi-
bition of this presentation (Fig. 5B). These inhibitors did not sig-
nificantly influence OVA257–264peptide presentation, suggesting
that PPV-VLP processing required the participation of some
proteases.
PPV-VLP processing requires proteasome activity
To evaluate whether the proteosomal complex also participates in
PPV-VLPs processing, CD11c?spleen cells were treated with two
20S proteasome inhibitors, lactacystin (50, 51) and LLnL (52).
OVA257–264presentation by DCs incubated with PPV-VLPs-OVA
FIGURE 4.
bated for 1 h with medium alone or in the presence of 0.2 nM primaquin (A) or 4 ?g/ml cycloheximide (B). Then 1.2 nM PPV-VLPs-OVA or 110 nM
OVA257–264was added, and DCs were incubated for 4 h in the continued presence of the drugs. After glutaraldehyde fixation, DCs were cocultured
overnight with 105cells/well of B3Z hybridoma. The presentation of OVA257–264epitope to B3Z cells was monitored by IL-2 production and expressed
as the percentage of [3H]thymidine incorporation compared with DCs incubated with PPV-VLPs-OVA or OVA257–264peptide in the absence of inhibitor.
Presentation of PPV-VLPs-OVA requires the neosynthesis of MHC class I molecules. Purified splenic DCs (105cells/well) were preincu-
2246CROSS-PRESENTATION OF VLPs BY DCs